Wyniki wyszukiwania - BazTech

1

Neuronowy generator wirtualnej zmiennej stanu w regulatorze prędkości układu dwumasowego

Stanisławski Radosław, Malarczyk Mateusz, Szrek Jarosław, Żychlewicz Mateusz, Kamiński Marcin

Przegląd Elektrotechniczny

|

2023

|

R. 99, nr 5

196--201

PL

W publikacji przedstawiona została modyfikacja klasycznego regulatora stanu, która zakłada wprowadzenie radialnej sieci neuronowej (Radial Basis Function Neural Network). Celem jest wygenerowanie sygnału, który zostanie wprowadzony do wektora zmiennych stanu sprzężeń zwrotnych. Obiektem regulacji jest napęd elektryczny z połączeniem sprężystym. W artykule uwzględniono opis teoretyczny proponowanego rozwiązania, a także zaprezentowano wyniki badań symulacyjnych struktury sterowania. Badania przeprowadzone dla rzeczywistego układu napędowego stanowią dodatkową weryfikację analizowanego regulatora stanu.

EN

In this paper, a state feedback controller enhanced by a Radial Basis Function Neural Network is presented. The main goal of the network is calculation of a virtual signal used in state vector and applied as feedback. The plant considered in the article is an electrical drive with a flexible joint. The mathematical description of the proposed control scheme and the numerical tests can be found in the manuscript. Experimental analysis is performed as an additional verification of the proposed state controller.

2

Radial basis function neural network and salp swarm algorithm for paddy leaf diseases classification in Thanjavur, Tamilnadu geographical region

Thirugnanasambandam Gayathri Devi, Rajkumar Ganesan, Srinivasan Anandan, Sudha Selvarajan

Acta Geophysica

|

2022

|

Vol. 70, no 6

2917--2932

EN

Farming is an essential sustenance for the progressive population. The development of our country depends on the farmers. Plants endure by many diseases due to environmental factors. So, the farmers need to detect plant diseases at an early stage for appreciable yield. In the beginning, the observing and examining plant disease are examined physically by the expertise in the farming field, which requires a considerable measure of work/ and requires over the top handling time. Now, machine learning concepts eliminate conventional protruding and time-consuming techniques. This paper focuses on a novel method for detecting and identifying paddy leaf diseases at the early stages in Thanjavur region using radial basis function neural network (RBFNN) classifier. Further, it is optimized with salp swarm algorithm (SSA) technique. The proposed method utilizes the data from the TNAU agritech portal, IRRI knowledge bank, UCI machine learning repository databases, which have healthy and diseased images. This work illustrates four categories (Bacterial Blast, Bacterial Blight, Leaf Tungro and Brown Spot) of infected paddy images along with the normal set of images. Initially the preprocessing is performed for the acquired images then K-means segmentation algorithm segregates the image. Gray level co-occurrence matrix extracts the Texture features from the segmented image and the RBFNN classifier performs the disease classification and improves the detection accuracy by optimizing the data using SSA. The investigational results of the proposed methodology exhibit the performance in terms of accuracy of disease detection is 98.47%. However, radial basis function neural network (RBFNN) achieves the diseases detection accuracy of 97.85% and support-vector machine (SVM) classifier achieves a disease detection accuracy of 97.07%. This paper proposes a method of paddy leaf disease recognition and classification using RBFNN and salp swarm algorithm. It also suggests and identifies an image analysis by framing a set of conditions for disease affected plants. The results show that the most satisfactory outcome can be gained to verify the yield of proposed methods with least effort.

3

Training RBF NN Using Sine-Cosine Algorithm for Sonar Target Classification

Wang Yixuan, Yuan LiPing, Khishe Mohammad, Moridi Alaveh, Mohammadzade Fallah

Archives of Acoustics

|

2020

|

Vol. 45, No. 4

753--764

EN

Radial basis function neural networks (RBF NNs) are one of the most useful tools in the classification of the sonar targets. Despite many abilities of RBF NNs, low accuracy in classification, entrapment in local minima, and slow convergence rate are disadvantages of these networks. In order to overcome these issues, the sine-cosine algorithm (SCA) has been used to train RBF NNs in this work. To evaluate the designed classifier, two benchmark underwater sonar classification problems were used. Also, an experimental underwater target classification was developed to practically evaluate the merits of the RBF-based classifier in dealing with high-dimensional real world problems. In order to have a comprehensive evaluation, the classifier is compared with the gradient descent (GD), gravitational search algorithm (GSA), genetic algorithm (GA), and Kalman filter (KF) algorithms in terms of entrapment in local minima, the accuracy of the classification, and the convergence rate. The results show that the proposed classifier provides a better performance than other compared classifiers as it classifies the sonar datasets 2.72% better than the best benchmark classifier, on average.

4

3D crosswell electromagnetic inversion based on radial basis function neural network

Fang Sinan, Zhang Zhansong, Chen Wei, Pan Heping, Peng Jun

Acta Geophysica

|

2020

|

Vol. 68, no. 3

711--721

EN

Crosswell electromagnetic (EM) method has fundamentally improved the horizontal detection ability of well logging and will become an increasingly promising approach for the secondary exploration of hydrocarbon reservoir. We applied orthogonal least squares (OLS) radial basis function neural network (RBFNN) based on improved Gram–Schmidt (G–S) procedure to three-dimensional (3D) crosswell EM inversion problems. In the inversion process of the simplifed crosswell model with single-grid conductivity anomalies and normal oil reservoir, compared the inversion results of other fve neural networks, OLS-RBFNN was proved to have the best global optimization ability and the fastest sample learning speed and the average inversion error of low conductivity anomalies model (4%) and oil reservoir model (9%) can meet the inversion requirements of crosswell EM method. Only the OLS-RBFNN could achieve ideal inversion results in the most concerned central area of crosswell model, and the inversion accuracy of this algorithm will be more outstanding when the model becomes more complex. Merely using the three-component time-domain crosswell EM data of two wells, the inversion of 3D medium conductivity in the crosswell dominant exploration area can be efectively realized through the nonlinear approximation of the OLS-RBFNN.

5

Soft computing-based technique as a predictive tool to estimate blast-induced ground vibration

Arthur Clement Kweku, Temeng Victor Amoako, Ziggah Yao Yevenyo

Journal of Sustainable Mining

|

2019

|

Vol. 18, iss. 4

287--296

EN

The safety of workers, the environment and the communities surrounding a mine are primary concerns for the mining industry. Therefore, implementing a blast-induced ground vibration monitoring system to monitor the vibrations emitted due to blasting operations is a logical approach that addresses these concerns. Empirical and soft computing models have been proposed to estimate blast-induced ground vibrations. This paper tests the efficiency of the Wavelet Neural Network (WNN). The motive is to ascertain whether the WNN can be used as an alternative to other widely used techniques. For the purpose of comparison, four empirical techniques (the Indian Standard, the United State Bureau of Mines, Ambrasey-Hendron, and Langefors and Kilhstrom) and four standard artificial neural networks of backpropagation (BPNN), radial basis (RBFNN), generalised regression (GRNN) and the group method of data handling (GMDH) were employed. According to the results obtained from the testing dataset, the WNN with a single hidden layer and three wavelons produced highly satisfactory and comparable results to the benchmark methods of BPNN and RBFNN. This was revealed in the statistical results where the tested WNN had minor deviations of approximately 0.0024 mm/s, 0.0035 mm/s, 0.0043 mm/s, 0.0099 and 0.0168 from the best performing model of BPNN when statistical indicators of Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Relative Root Mean Square Error (RRMSE), Correlation Coefficient (R) and Coefficient of determination (R2) were considered.

6

2D Cadastral Coordinate Transformation using extreme learning machine technique

Ziggah Y. Y., Issaka Y., Laari P. B., Hui Z.

Geodesy and Cartography

|

2018

|

Vol. 67, no. 2

321--343

EN

Land surveyors, photogrammetrists, remote sensing engineers and professionals in the Earth sciences are often faced with the task of transferring coordinates from one geodetic datum into another to serve their desired purpose. The essence is to create compatibility between data related to different geodetic reference frames for geospatial applications. Strictly speaking, conventional techniques of conformal, affine and projective transformation models are mostly used to accomplish such task. With developing countries like Ghana where there is no immediate plans to establish geocentric datum and still rely on the astro-geodetic datums as it national mapping reference surface, there is the urgent need to explore the suitability of other transformation methods. In this study, an effort has been made to explore the proficiency of the Extreme Learning Machine (ELM) as a novel alternative coordinate transformation method. The proposed ELM approach was applied to data found in the Ghana geodetic reference network. The ELM transformation result has been analysed and compared with benchmark methods of backpropagation neural network (BPNN), radial basis function neural network (RBFNN), two-dimensional (2D) affine and 2D conformal. The overall study results indicate that the ELM can produce comparable transformation results to the widely used BPNN and RBFNN, but better than the 2D affine and 2D conformal. The results produced by ELM has demonstrated it as a promising tool for coordinate transformation in Ghana.

7

Zastosowanie radialnej sieci neuronowej jako adaptacyjnego regulatora prędkości napędu dwumasowego

Kamiński M., Orłowska-Kowalska T.

Przegląd Elektrotechniczny

|

2014

|

R. 90, nr 5

41--44

PL

Artykuł prezentuje zastosowanie sieci neuronowej adaptowanej on-line w pętli regulacji prędkości układu napędowego z połączeniem sprężystym. W algorytmie zastosowano model neuronowy z radialnymi funkcjami aktywacji. W celu aktualizacji wartości wag oraz centrów regulatora adaptacyjnego zastosowano algorytm gradientowy. W tej części obliczeń poprawki dla aktualnych parametrów regulatora neuronowego są mnożone przez stałe determinujące stopień oddziaływania zastosowanej metody adaptacji. Charakterystycznym rozwiązaniem, prezentowanym w niniejszym artykule, jest zastosowanie modelu rozmytego w celu wyznaczenia wspomnianych współczynników skalujących. Zaprojektowany regulator został przetestowany w symulacjach oraz wykonano badania eksperymentalne z wykorzystaniem procesora sygnałowego karty dSPACE1103.Uzyskane wyniki prezentują precyzję sterowania analizowanego regulatora neuronowego oraz jego odporność na zmiany parametrów obiektu.

EN

Article presents application of neural network trained on-line in speed control loop of electrical drive with elastic connection. In algorithm neural model with radial activation function was implemented. For updates of weights and centers of adaptive controller gradient method was used. In this part of calculations correction values for adaptive controller are calibrated. Specific solution, described in paper, is application of fuzzy model for determination of scaling coefficients. Designed controller was tested in simulations and then experiment was prepared (using dSPACE1103 card). Achieved results present high precision of control and also robustness against changes of the drive parameters.

8

Analysis of alkaloids in pharmaceutical preparations containing Kushen by capillary electrophoresis with application of experimental design and a quantitative structure-property relationship approach

Wen Y., Liu H., Tian L, Han P., Luan F.

Acta Chromatographica

|

2010

|

Vol. 22, no. 3

445--457

EN

A simple and rapid capillary electrophoretic procedure for analysis of matrine and oxymatrine in Kushen medicinal preparations has been developed and optimized. Orthogonal design was used to optimize the separation and detection conditions for the two active components. Phosphate concentration, applied potential, organic modifier content, and buffer pH were selected as variable conditions. The optimized background electrolyte contained 70 mM sodium dihydrogen phosphate and 30% acetonitrile at pH 5.5; the separation potential was 20 kV. Each analysis was complete within 5 min. Regression equations revealed linear relationships ( r > 0.999) between peak area and amount for each component. The detection limits were 1.29 μg mL -1 for matrine and 1.48 μg mL -1 for oxymatrine. The levels of the two active compounds in two kinds of traditional Chinese medicinal preparation were easily determined with recoveries of 96.57–106.26%. In addition, multiple linear regression and a non-linear model using a radial basis function neural network approach were constructed for prediction of the migration time of oxymatrine. The predicted results were in good agreement with the experimental values, indicating that a radial basis function neural network is a potential means of prediction of separation time in capillary electrophoresis.